Part Number: AM5728
Tool/software: Linux
Hi. We have developed linux( Three display and Video processing and rendering with 3D GPU) and rtos( audio processing engine( audio effect, mixing, reverb with various FFT ).
PCM is shared on CMEM between linux and rtos.
Sound of DSP(C66x) is very good, when linux only run on multi audio player and don't run on Video processing and rendering with 3D GPU.
but When Linux running on multi audio player, Three display, Video processing and rendering with 3D GPU, Sound of DSP(C66x) with McASP is broken.
and, we test it as below .
- When Linux running on multi audio player, Three display and Video processing, Sound of DSP(C66x) with McASP is good. sometimes Sound have glitch.
- When Linux running on multi audio player, Three display, Video processing and rendering with 3D GPU, Sound of DSP(C66x) with McASP is broken.
we wonder that Memory bandwidth enough.
./glsdkstatcoll -f config.ini
EMIF1 cmd_pend 86% data 51%
EMIF2 cmd_pend 83% data 50%
How does it make good sound with linux full use many IPs(DSS, 3DGPU, 2DGPU, IVA-HD) ?
If DSP have highest prioriy on PEG of DMM and other many IPs(DSS, 3DGPU, 2DGPU, IVA-HD) have lowest prioriy, Is it sound pretty good ?
- Is it right, we want how to settings. (ex. IVA 4-bit ConnID is 3 at Table 14-10(ConnID Values) of TRM, but IVA 4-bit ConnID is E at e3e sites, and then we don't know proper ConnID ).
If you have other solution, please kindly answer me.
Thanks a lot.